The five lighting spec mistakes we see every week

1. Specifying CCT without CRI

3000K is a temperature, not a quality. A 3000K fixture at CRI 80 makes a beautiful walnut floor look grey. Always specify CRI ≥ 95 for any material-critical zone.

This is the single most common mistake we see, and it persists because CCT (correlated colour temperature) is the number everyone learns first and CRI (colour rendering index) is the one that actually determines whether materials look right under the light. A fixture can hit exactly the right warm 3000K "feel" while still rendering colours poorly — CRI 80 fixtures are common, cheap, and genuinely fine for back-of-house or storage areas, but disastrous anywhere a client is meant to appreciate a material finish. We've watched a beautifully specified walnut floor get value-engineered down to a CRI 80 fixture at the eleventh hour to save cost, and the client complaint that followed — "why does the floor look muddy?" — cost more in remedial lighting work than the saving was worth.

2. Forgetting beam angle on accent lights

A 36° beam on a 4m ceiling washes a wall. A 10° beam picks out an artwork. They cost the same. Spec the right one.

Beam angle is one of those specs that gets treated as an afterthought because, unlike CRI or CCT, it doesn't show up as a headline number on most spec sheets — it's buried in a datasheet PDF. The practical impact is enormous. A wide beam angle floods an area evenly, appropriate for general ambient wash. A narrow beam angle creates a distinct pool of light with a defined edge, appropriate for highlighting a specific object. Using the wrong one doesn't just look slightly off — it fundamentally changes what the light is doing in the room. We've seen entire gallery-wall lighting schemes redone post-installation because someone specified 36° fixtures throughout when half the fixtures were meant to be picking out individual artworks with a tight 10° beam.

3. Mixing driver brands

One project, one driver family. Mixed drivers = mixed dimming curves = flicker at 20%. The client always notices. Always.

LED drivers control how a fixture dims, and different manufacturers implement dimming curves — the relationship between the dimmer's position and the actual light output — differently. Mix drivers from two brands on the same dimming circuit and the fixtures don't dim in sync; one bank of lights hits 20% brightness smoothly while another flickers or steps unevenly at the same dimmer setting. This is invisible on a spec sheet and completely obvious in a finished, dimmed room. It typically happens when a project runs out of a specific driver mid-installation and a site team substitutes "an equivalent" from whatever's on hand rather than escalating back to the design team. The fix after the fact means re-wiring and re-commissioning an entire dimming zone — expensive and disruptive on a space that's otherwise finished.

4. Ignoring glare cut-off

Anything below 30° UGR in a hospitality lounge is a complaint waiting to happen. Deep baffles, honeycomb louvres, or recessed pinholes.

UGR (Unified Glare Rating) measures how much a fixture's light source is directly visible and uncomfortable to someone seated or standing in the space, rather than how much useful light it produces. A fixture can deliver exactly the right lux level and colour rendering while still being genuinely unpleasant to sit under if the light source itself is too visible at a typical eye-line angle. Hospitality lounges are particularly exposed to this because guests spend long periods seated, at a consistent eye height, often looking toward the ceiling or across the room — exactly the conditions where a poorly shielded fixture becomes a persistent low-grade irritant that guests can't quite articulate but that shows up in reviews as "the lighting felt harsh."

5. Under-spec'd drivers

A 90W fixture on a 100W driver runs hot and dies in 18 months. Spec drivers at 120% of fixture load, minimum.

This is a cost-cutting decision that almost never gets flagged until well after handover, because a driver running close to its rated capacity works fine initially — it just runs hotter than it should, and that heat steadily degrades the driver's components over months rather than causing an immediate failure. By the time fixtures start failing in year two, nobody connects it back to a driver spec decision made during procurement. The fix is simple and cheap at spec stage — build in the 120% headroom — and expensive and disruptive after the fact, when failed drivers mean re-opening finished ceilings to replace them.

Two more that don't make the classic list but should

Ignoring dimming compatibility between the driver and the building's control system. A driver that dims beautifully on a bench test can behave erratically once connected to a building-wide DALI or 0–10V control system it wasn't tested against. Confirm driver-to-control-system compatibility in writing before ordering at scale, not after the first fixture is installed and behaving oddly.

Under-speccing colour consistency (MacAdam ellipse) across a large run of the same fixture. Order 200 identical downlights from the same product line and, without a tight MacAdam ellipse specification (SDCM 3 or better, ideally), you can end up with visibly different white points across the ceiling — some fixtures reading slightly warmer, some slightly cooler, in a way that's subtle in isolation and obvious once you're standing under a hundred of them at once.

Building this into a checklist

Every one of these mistakes shares a common root: they're invisible on a basic spec sheet and only become obvious once the space is built and occupied. The fix isn't more experience alone — it's a specification checklist that forces CRI, beam angle, driver compatibility, UGR and dimming behaviour to be confirmed in writing before an order goes out, not assumed because "that's what we always spec." The projects that avoid these five mistakes aren't the ones with the most talented lighting designer. They're the ones with the most boring, thorough checklist.